一种融合情感规则与机器学习的情感分类方法

A method of emotion classification which combines emotion rules and machine learning

针对评论型长文本的情感倾向性问题,提出了一种融合情感规则与机器学习的分类方法.基于情感规则得出评论的情感得分,该方法将文本分解为一组子句,以词汇为基本颗粒进行分数计算,得出最佳位置权重系数.同时,不同类型句式共归纳出4类关联词与之对应.将所得权重系数与关联词得分相结合,总结出情感计算公式.然后将所得情感得分作为特征融合到机器学习分类器的输入矩阵中,构造最优情感分类器.实验所得最优分类器准确率为0.979,高于同类算法.

A classification method combining emotional rules and machine learning is proposed to solve the problem of emotional orientation of long critical texts.First of all,the emotional score of the comment is obtained based on the emotional rules,the method refines the text into a set of clauses,with vocabulary as the basic particle scores calculated,it is concluded that the best position weight coefficient.Meanwhile,there are four types of related words corresponding to different types of sentence patterns.Combining the weight coefficient with the score of related words,the formula of emotion calculation is summarized.Then,the obtained emotion score is integrated into the input matrix of machine learning classifier to construct the optimal emotion classifier.The accuracy of the optimal classifier is 0.979,higher than the similar algorithm.